T BING 1201205 Chapter3

(1)

Siti Nuraeni Muhtar_{, 2014}

STUDENTS’ FOREIGN LANGUAGE CLASSROOM ANXIETY (FLCA)

IN EFL CLASSROOM: IT’S LEVELS, SOURCES AND COPING STRATEGIES

Universitas Pendidikan Indonesia | repository.upi.edu | perpustakaan.upi.edu

34 CHAPTER III

METHODOLOGY

This chapter describes the methodology used to collect and analyze data in

this study. There are six sections in this chapter. In the first section, aims of the

study are displayed in the introduction section to show readers the research

problems addressed in this study. Second, research design is discussed to inform

research method and approach applied in this study. Next, data collection is

described to show research site, subjects of study, steps of data collection and

techniques used in collecting the data. Finally, data analysis is explained to report

how the collected data are analyzed step by step to answer research questions.

3.1 Introduction

It was acknowledged in chapter one that this research mainly aims to

report students‟ anxiety in English foreign language classroom. To fulfill the aim,

three research questions are proposed; the first deals with language anxiety level

that students‟ experienced in EFL classroom, the second relates with sources of

students‟ language anxiety and the last investigates the coping strategies students‟

used for their foreign language anxiety. The following sections elaborate the

methodology used in collecting and analyzing the data to answer the proposed

research questions.

3.2 Research Design

Based on the categorization of research design from Nunan and Bailey

(2009), this study can be classified as a case study research because the case in

this study has physical boundary (this study was conducted in a certain State

(2)

35

behavior which had a beginning and an end) and bounded instance (this study

focused on one classroom). This study was also conducted in naturally occurring

situation and context without manipulating variables (p.162). Hence, the result is

centered one description, inference and interpretation.

Three types of case study research designs includes exploratory,

explanatory and descriptive. (Hancock & Algozzine, 2006). This study belongs to

descriptive designs which attempt to present a complete description of a

phenomenon within its context. (p.33)

This study applied a qualitative approach by giving questionnaires and

taking transcripts of observation and interview to find students‟ language anxiety

in EFL classroom. To make the description clearer, some qualitative data were

quantified to show the number, percentage and distribution of students‟ language

anxiety. In this case, Nunan and Bailey (2009) state:

All qualitative data can be quantified in some way. In other words, things can be counted in qualitative data. In fact, there is almost no limit to the things that can be counted in qualitative data sets. Consider a lesson

transcript….(Nunan & Bailey, 2009, p.414)

The description of research design and data analysis hopefully can give a

clear picture on how this study was carried out to answer the three research

questions mentioned above.

3.3 Data Collection

This part answers who were taken as subjects of study and where, when

and how the data were collected:

3.3.1 Research Site

This study was conducted in an EFL classroom in one of State University

(3)

36

English as foreign language (EFL). There were three classes (class A, class B, and

class C). The class was divided based on the placement test scores. Class A

belongs to students who get the highest scores, class B belongs to moderate

students, and class C belongs to those who get the lowest scores.

This course was chosen as research site for several reasons. The first

reason was the English course serves as a typical case (Nunan & Bailey, 2009) to

find out information from a specific case to fill in the gap of previous research. In

this case, the site fit with the researcher‟s intention to investigate students‟

language anxiety in EFL classroom. The program in this course is very fit with

this study which related to EFL learning.

The second reason was the researcher‟s convenience (Patton, 1980, in

Alwasilah, 2002) to conduct research in chosen research site due to its support and

accessibility. The research was fully supported by the director and the lecturers

since the researcher is one of the lecturer there and the research was expected to

improve the teaching learning activity in the classroom. In addition, the research

site was fully accessible. In this case, the researcher was fully permitted and

supported to conduct research in the classroom.

3.3.2 Subjects of the Study

After several considerations, class A was chosen. This class was taken due

to several reasons. The students in class A was more expressive and corporative

comparing to other classes. The class consists of 35 students who got the best

scores in placement test. The students in the class were also considered as active

students who usually take part in class interaction.

Classroom interaction usually lasted in 90 minutes. However, the

interaction that was transcribed only focused on those students who were

(4)

37

same required classes offered by the university to students of different ability

levels that would train them in the skills of English listening, speaking, reading,

and writing. It should be noted that although the participants were homogeneously

grouped for English classes, when the data collection procedure was undertaken,

they were only at their second semester.

3.3.3 Steps of Data Collection

Data collection for the present study was conducted in four weeks. The

following table is the schedule of data collection for this study:

Table 3.1

The Schedule of Data Collection

Monday, March 3rd, 2014 90 minutes Video – audio recording, field note

Wednesday, March 5th, 2014 90 minutes Taking questionnaire data

Monday, March 10th, 2014 90 minutes Video – audio recording, field note

Wednesday, March 12th, 2014 30 minutes Taking interview data (2 participants)

Monday, March 17th, 2014 30 minutes Taking interview data (2 participants)

Wednesday, March 19th, 2014 30 minutes Taking interview data (2 participants)

Monday, March 24th, 2014 90 minutes Video – audio recording, field note

Observation in the class was conducted for three meetings. Those three

meetings covered listening-speaking class, reading class and grammar class. The

interviews were conducted before the class started.

(5)

38

This section describes the procedures utilized to collect the data. Broadly

speaking, there are three techniques for collecting qualitative data: observation,

interview, and document review (Best & Kahan, 2005). There was also three data

collection techniques employed in this study. Questionnaires, observation and

semi-structured interview. The three techniques were used to corroborate the

findings to answer the research questions. By employing three different

techniques, this study is expected to provide triangulated findings and analysis on

students‟ language anxiety in EFL classroom.

First, the participants were observed in their EFL classroom for a period of

four weeks. Informal interviews were conducted as frequently as possible in order

to acquire additional information useful in checking the accuracy of the

impressions obtained during the observation period. In addition, semi-structured

interviews were conducted in the period of observation with more detailed

questions. Field notes were written during and immediately after each

observation, as suggested by Emerson, Fretz, and Shaw (1995). The interviews

were conducted in native languages, Indonesian in the case of the student

participants. All semi-structured interviews were tape-recorded and transcribed.

3.3.4.1 Observation

The purpose of observation is to collect the qualitative data that are helpful

in answering the research questions of this study - detailed notations of behaviors,

events and contexts related to language anxiety. As explained before in the

discussion of the merits of qualitative research, observation has several

advantages in such a task (Patton, 1980, as cited in Alwasilah, 2002) make a better

understanding of the context, within which students react and cope with their

language anxiety; 2) allow for inductive approach, thereby making it less

(6)

39

on sensitive topics, such as interpersonal competition among certain students, that

they may not be willing to talk about in an interview; 4) provide the resources

otherwise unattainable - observer‟s impression, feeling, reflection and

introspection. These can be part of meaningful data to allow for understanding and

interpreting the events and behaviors with a holistic perspective.

To achieve those purposes, the classroom interaction was observed

electronically by recording the interaction using audio and video recording

(Nunan & Bailey, 2009). Besides, the researcher presented there to make sure the

recorders worked well and jot down some important things regarding the focus of

this study. During three meetings (90 minutes per meeting) of lesson were

recorded using a video camera and a digital audio recorder. Before the students

entered the class, the camera was turned on and put on the three port behind the

students. to make sure the voice could be well recorded; an audio recorder was

turned on and put on teachers‟ table in front of the class. By setting up a camera at

the back side and an audio recorder in the front side of the classroom, it was

hoped that all interactions were well documented.

In observing classroom interaction, the researcher took a role as a

non-participant observer. The researcher did not take a part in planning the lesson nor

in teaching and learning process. She only put the video and audio recorder in a

hidden place in the classroom. The lecturer was informed that the researcher was

going to record the classroom interaction without specifically mentioning what

features of interaction were going to be investigated.

Even through a very attentive observation, it is hardly possible to capture

everything happening in the classroom. Therefore, it is necessary to focus on

certain kinds of activities and events that provide the particular information and

insights helpful for answering the research questions of this study. This strategy

(7)

40

behaviors relevant to the purpose of this study, but also makes the observation

work manageable (Patton, 1987). Conforming to this strategy, the focus of the

observation (i.e., what to observe) in this study was made on the following

elements, as suggested by Merriam (1998, as cited in Cha, 2006):

1) The physical setting

A close observation on physical environment (e.g., classroom layout) is

indispensable in this study, since it can affect students‟ feelings and

behaviors. For instance, a particular classroom layout facilitating group

work, can promote a higher level of student to student interactions during

class and thus their sense of camaraderie, which has been found to render a

diminishing effect on language anxiety (Young, 1990; Samimy & Rardin,

1994)).

2) The participants

This element of observation will identify who are in the classroom, that is,

demographic characteristics of participants, including not only main

participants but also their classmates and teachers. It is important to

understand the demographic characteristics of all participants, for they

may yield useful clues to appreciating the mechanism of language anxiety,

which is otherwise difficult to grasp.

3) Activities and interaction

The observation was made on specific activities and behaviors of all

participants relevant to language anxiety, such as the competitiveness,

level and types of interactions, sequence of activities. For example, the

researcher tried to find out the pattern of teacher‟s manner of error

correction, and to see whether or not student‟s making a mistake in

English is followed by giggling or lending a hand.

(8)

41

The content of conversations among students as well as with teachers was

noted. Their conversations were directly quoted, paraphrased or

summarized in the field notes, as needed to answer the research questions.

In addition, I also noted silence and nonverbal behaviors, such as

avoidance and withdrawal that have important bearing of language

anxiety.

5) Subtle factors

The researcher looked for less obvious but relevant behaviors to language

anxiety that contain symbolic and connotative meanings such as biting

nails, making face, sweating, gazing, the way of raising hand, etc. More

importantly, the researcher was equally attentive to „what does not

happen,‟ in other words, “absence of occurrence” (Patton, 1987; p. 91).

6) My own behaviors

Merriam (1998) points out that the thoughts, comments, roles of the

observer are as important as those of participants, since like the latter, the

former also affect what is being observed. Put differently, the observation

is not merely a close look on the scene with perfect neutrality, but

necessarily involves a cognitive process that cannot totally exclude

subjectivity of the observer (Fraenkel & Wallen, 2003). Therefore, what

the observer is thinking about what is going on in observation is necessary

to be noted, for it can be used later to have as wide a perspective as

possible in deciphering the sequences and patterns of events and activities

recorded in the field notes.

3.3.4.2 Semi-Structured Interview

Interviews most often used in qualitative research can be divided into three

(9)

42

structured or standardized) interview (Fraenkel & Wallen, 2003; Merriam, 1998).

These types of interview are distinguished by the amount of structure involved.

The informal interviews of this study do not have any predetermined question

topics or sequence of questions or any particular forms of questioning. Questions

would come forward from the immediate context as the conversation unfolds, and

would be asked in the natural course of conversation.

Although informal interview is prone to produce less systematic and

comprehensive data, it allows the researcher to be responsive to individual

differences of participants and to enhance the salience and relevance of questions

(Fraenkel & Wallen, 2003). The second type of interview used in this study was a

semi-structured interview. The semi-structured interview has a fairly open

framework, thereby allowing a great deal of flexibility in the kind of information

that interviewees choose to give (Patton, 1990). Furthermore, this type of

interview offers “sufficient flexibility to probe some aspects in depth and, where

necessary, to let the respondent lead in much the same way as in open interview”

(Heigham and Croker, 2009, p. 186, as cited in Cha, 2006). In other words, unlike

formal (standardized) interview, where detailed questions are prepared in advance,

semi-structured interview has the questionnaire framework that starts with more

general questions or topics. As such, semi-structured interview is formulated

ahead of time only in the sense that some form of interview guide is prepared

beforehand and provides a framework for the interview. Similarly, a large part of

questions in semi-structured interview are not formulated and phrased in advance,

but created during the interview, enabling both the interviewer and the

interviewees the flexibility to probe for details or discuss issues. This facilitates

gaining the information the interviewer needs without overly influencing the

participants' responses. This flexibility during the interview is important in

(10)

43

hand while bringing up information that the researcher might not anticipate,

potentially shedding a whole new insight onto a problem (Merriam, 1998;

Fraenkel & Wallen, 2003). In summary, semi-structured interview can be

characterized as a relatively guided and relaxed discussion based around a

predetermined topic.

Furthermore, To make the interview like a guided conversation, broad

questions were asked so as not to constrain the conversation. As the

semi-structured interview, moreover, the set of questions had been prepared in advance,

but they were open enough to allow the interviewees to express their thoughts,

feelings, or opinions freely throughout the interview. And questions were

generally simple with a logical sequence to help the interview move naturally. For

an example, first, a broad question was asked such as “How was speaking English

in EFL class?” If the answer to the above question is “difficult,” then, new questions came forward as a result of the answer, such as “Why is that so?”

The complete set of interview questions had not been formulated in

advance, since many of relevant interview questions emerged during observation.

They were formulated fully as observation showed issues, behaviors, events and

so on that formed the basis of interview questions. The language used in

interviews with participants was Bahasa Indonesia. Since the researcher

conducted interviews with the participants in Bahasa Indonesia, their native

language, it was possible for the researcher to catch the linguistic nuances and

cultural connotations behind their overt expressions. It appeared that the use of

native language in interviews has compensated the difficulties often arouse by the

interviewees to express themselves clearly.

To conduct a semi-structured interview, the interviewer focused on the

analysis of students‟ language anxiety using Foreign Language Classroom

(11)

44

exhibited high level of anxiety were interviewed according to a set of

semi-structured questions to find out their source of language anxiety and their coping

strategies. In addition, two EFL students who exhibited moderate level of anxiety

and two EFL students who exhibited low level anxiety were interviewed also. The

time required for the interview ranged from 10 – 15 minutes.

The detail schedules for the interview can be seen as follows.

Table 3.2

The Detail Schedule of the Interview

No Student/Subject Date Duration

1. Student 1 (high anxiety student) Wednesday, March 12th, 2014 9:54 minutes

2. Student 2 (high anxiety student) Wednesday, March 12th, 2014 10:13 minutes

3. Student 3 (moderate anxiety student) Monday, March 17th, 2014 12:05 minutes

4 . Student 4 (moderate anxiety student) Monday, March 17th, 2014 9:30 minutes

5. Student 5 (low anxiety student) Wednesday, March 19th, 2014 10:11 minutes

6. Student 6 (low anxiety student) Wednesday, March 19th, 2014 11:32 minutes

Henceforth, the interview data were interpreted to answer research problems and

compared to the theory underpinning the study (Emilia, 2009, p. 197).

3.3.4.3 Questionnaires

The 33-question Foreign Language Classroom Anxiety Scale (FLCAS)

developed by Horwitz et al. (1986) was translated into Bahasa Indonesia by the

researcher and used to measure learner anxiety in this study. There were five

response choices for each statement: strongly agree (SA), agree (A), neutral (N),

(12)

45

different point per item (1 = strongly disagree; 2 = disagree; 3 = neither agree nor

disagree ; 4 = agree; 5 = strongly agree). A good instrument (whether test or

non-test) must be valid and reliable (Sugiyono, 2011,. P.169). This is intended to

collect the valid and reliable data. Sugiyono (2011,p. 168) notes that by using the

valid and reliable instruments in collecting data, it is expected that the result of the

study is also valid and reliable. The Indonesian version of this instrument is highly

reliable, with a coefficient alpha of .96. In terms of validity, basically, validity is

divided into three types: content validity, criterion-related validity, and construct

validity (Hatch and Farhady, 1982, p.251). Since the instrument used in this study

was a non-test instrument, which was FLCA, according to Sugiyono (2011,

p.170), to be stated as a valid instrument, it only needed to fulfill construct

validity.

Dealing with construct validity, Hadi (cited in Sugiyono, 2011, p.170)

treats construct validity like logical validity or validity by definition. Thus, an

instrument can be stated for having construct validity if it can be used to measure

the phenomenon as its definition is. Further, to utter an appropriate definition, the

researcher needs several theories to consult with. In this case, Hadi states that

when the theories used to utter a definition are already appropriate, the result of

measurement using the instrument which is based on those theories is already able

to be stated as a valid result.

To test the construct validity, the judgment from experts was employed in

this study. For expert judgment, at least three doctorate experts are needed

(Sugiyono, 2011, p. 172). Hence, in this study, the judgment was done by two

doctorate lecturers from English Education (who concern in Second Language

Acquisition) and a doctorate lecturer from Educational Psychology.

(13)

46

Data analysis in this study focused on data collected above: data from

observation, interview and questionnaire. The observation data result was

transcribed, coded, categorized and analyzed. Meanwhile, the result of interview

was coded and analyzed. Then the data from questionnaire was counted, displayed

and analyzed. At last, the analysis of each data collection was synthesized and

discussed to answer the research questions. The following parts explain the

process of analysis of each collected data to fulfill the aims of study.

3.4.1 Analyzing the Data from Observation

The data from observation (audio and video recording) were transcribed,

coded, categorized, described and analyzed to reveal the different types of

language anxiety source and coping strategies.

The process of analysis started with transcribing the data. In this case, the

transcription focused on the following elements, as suggested by Merriam (1998,

as cited in Cha, 2006):

1) The physical setting

2) The participants

3) Activities and interaction

4) Conversation

5) Subtle factors

6) My own behaviors

After the recording of interaction were transcribed, the data were coded

and categorized by adapting the categorization of language anxiety coping

strategies by Erhman (1996) to answer the third research question collaborating

with interview result.

(14)

47

The data from interview were transcribed and analyzed qualitatively. The

interview data were analyzed to collaborate and confirm the findings from

observation and questionnaire in answering the second and third research

question. In analyzing the data, the researcher carried out several steps. First, data

from interview were transcribed. Second, the transcribed data were coded by

following Alwasilah‟s coding strategy (2002, p.232) by categorizing the data

based on the responses given by the teacher. Third, the result of coding was

condensed to confirm or contradict the findings obtained from observation and

questionnaire.

In supporting the findings from observation, the findings from interview

were cited in the analysis following Creswell‟s (2003, p.197) suggestion to use the

wordings from participants to give a detailed descriptive portrait. The wordings

from interview were also compared and contrasted with the data from observation,

theories and previous studies on language anxiety.

Considering the guidelines for interview data, in categorizing this result,

the researcher used Horwitz categories for language anxiety sources

(communication apprehension, test anxiety and fear of negative evaluation) in

order to answer the second research question. Besides, Ehrman‟s (1996) defense

mechanism categories for students‟ coping strategies of language anxiety (flight

behavior, aggressive behavior, group manipulation behavior, and compromise

behavior) were used to answer the third research question.

3.4.3 Analyzing the Data from Questionnaire

The data from questionnaire, which is adapted from Foreign Language

Classroom Anxiety Scale (FLCAS) by Horwitz, was analyzed quantitatively.

FLCAS consists of two kinds of statement which are positive and negative. The

(15)

48

“Strongly Disagree”. They are 9 positive statements in the questionnaire which are

number 2, 5, 8, 11, 14, 18, 22, 28, and 32. While, the negative statements are in

number 1, 3, 4, 6, 7, 9, 10, 12, 13, 15, 16, 17, 19, 20, 21, 23, 24, 25, 26, 27, 29, 30,

31, and 33. The data were calculated manually with the range of score started

from 61 to 139. After the each score from students‟ results of FLCAS gained, it

was categorized into some levels of anxiety started from “High Anxious”,

“Moderate Anxious”, and “Low Anxious”.

The Likert‟s scoring scale table to measure students‟ anxiety level using FLCAS is

shown below:

Table 3.3

Likert‟s Scoring Table adopted from Horwitz et al. (1986)

Statement

Scoring

Strongly

Agree Agree

Neither

Agree nor

Disagree

Strongly

Disagree

Positive 1 2 3 4 5

Negative 5 4 3 2 1

The following table is the score for each anxiety level perceived by students:

Table 3.4

FLCAS Anxiety Scale

Range Level

(16)

49

percentage. Furthermore, to calculate the mean of anxiety level to each item, first

multiplied a score of each answer by its percentage, divided the calculated score

by the sum of percentage. The means of every item can clarify the overall level of

students‟ language anxiety.

3.5 Validity and Reliability

In maintaining validity, the researcher did several things:

a. Methods Triangulation: Multiple methods were used in collecting the data

including observation, interview and questionnaire.

b. Theory triangulation: Various theories were brought to bear in this study

including interaction, output and noticing theories.

c. Member validation/member checking: It involved asking the participants

after the questionnaire to confirm their answers and interpretation.

In maintaining reliability of this study, the researcher applied interceding

process, following Nunan&Bailey (2009) who suggest researchers to conduct

intercoder agreement with the following steps:

One way to sort out this problem is to determine intercoder agreement-an index of the consistency with which different people categorize the same

In applying intercoder agreement, the researcher carried out several steps.

(17)

50

the researcher‟s assumption of her linguistic ability. Second, the partner was

trained to code the data base on detail descriptors. Third, the researcher and the

intercoder coded 10 % of data together as an internalization of agreement toward

the descriptor. Fourth, the researcher asked a partner to code 90 % of data (see

Sheen, 2004) and looked at the result whether agreement reached more or less

than 85 % after the coding process was finished; the results were compared to

look at the level of agreement. As a result, the agreement for learners‟ anxiety

codification reached 95 % while the agreement for interview codification reached

86 %. The differences were resolved by discussion and modification of descriptor.

3.6 Concluding Remarks

This chapter had drawn the qualitative study, particularly the case study

design, as the methodology of the study. Therefore, the data were collected by

means of the questionnaire analysis, the observation analysis and the interview.

Consequently, all the data obtained were analyzed qualitatively to be compared